The postdoc will be based at the CELIA laboratory which develops studies on different patterns of inertial fusion by laser. In order to optimize the implosion of the target, the laser pulse is shaped spatially and temporally, in particular by a pre-pulse of a hundred picoseconds and intensity of a few hundred TW /cm2. However, the latter introduces spatial inhomogeneities to the surface and volume of the target, amplified by the initial solid behavior of matter. These fingerprints generated by the pre-pulse will degrade the symmetry of the target during its implosion, and therefore decrease the effectiveness of inertial confinement. At present, most models assume a plasma state from the beginning of the interaction, and are thus unable to account for certain experimental observations. To overcome this lack, we have just developed an original multi-physics simulation tool that includes the phase transition of a homogeneous material induced by the laser. In order to mitigate the laser imprint effect, a polystyrene foam (heterogeneous material) can be deposited on the surface of the target. The multiple optical reflections in the foam smooth the spatial profile of laser intensity, thus reducing absorption inhomogeneities. In order to reduce the influence of the laser fingerprint, the post-doctoral fellowship will aim to develop a microscopic model describing the evolution of the optical response of a foam during the solid-to-plasma transition. The first step of the work will be to couple the Helmholtz equation (describing laser propagation) to a solid transition model-plasma, and to study the influence of parameters. The second step will be to use an artificial intelligence algorithm (neural network) to optimize the optical response of the foam.